eAppendix: Supplemental Material
This supplement includes additional references relevant to the points made and additional figures that support and clarify the conclusions.
Additional References
S1. Brunekreef B. Air pollution and life expectancy: is there a relation? Occup Environ Med. 1997;4:781-784. .
S2. Englert N. Time-series Analyses and cohort studies to investigate relationships between particulate matter and mortality – Two approaches to one endpoint. J. Environ Med. 1999;1:291-296
S3. Centers for Disease Control and Prevention (CDC). Annual smoking-attributable mortality, years of potential life lost, and economic costs--United States, 1995-1999. MMWR Morb Mortal Wkly Rep. 2002 Apr 12;51(14):300-3.
S4. Giesinger I, Goldblatt P, Howden-Chapman P, Marmot M, Kuh D, Brunner E Association of socioeconomic position with smoking and mortality: the contribution of early life circumstances in the 1946 birth cohort. J Epidemiol Community Health. 2014 Mar 1;68(3):275-9.
S5. Singh GK, Siahpush M. Widening rural-urban disparities in life expectancy, U.S.., 1969-2009. Am J Prev Med. 2014 Feb;46(2):e19-29.
S6. Vermeulen R, Silverman DT, Garshick E, Vlaanderen J, Portengen L, Steenland K.
Exposure-Response Estimates for Diesel Engine Exhaust and Lung Cancer Mortality
Based on Data from Three Occupational Cohorts. Environ Health Perspect. 2014 122:172-177.
S7. Guo P, Yokoyama K, Suenaga M, Kida H. Mortality and life expectancy of Yokkaichi asthma patients, Japan: late effects of air pollution in 1960-70s. Environ Health.2008;7:8.
S8. Pope CA 3rd, Ezzati M, Dockery DW. Fine-particulate air pollution and life expectancy in the United States. N Engl J Med. 2009; 360:376-386. Also, Lipfert FW. Air pollution and life expectancy. N Engl J Med. 2009;360:2033 and author reply 2033-2044.
S9. Chen Y, Ebenstein A, Greenstone M, Li H. Evidence on the impact of sustained exposure to air pollution on life expectancy from China's Huai River policy. Proc Natl Acad Sci U S A. 2013;110:12936-12941.
S10. Escobedo LG, Peddicord JP. Smoking prevalence in US birth cohorts: the influence of gender and education. Am J Public Health. 1996 Feb;86(2):231-6.
S11 Guo Y, Li S, Tian Z, Pan X, Zhang J, Williams G. The burden of air pollution on years of life lost in Beijing, China, 2004-08: retrospective regression analysis of daily deaths. BMJ. 2013 Dec 9;347:f17139
S12. Hedley AJ, Wong CM, Thach TQ, Ma S, Lam TH, Anderson HR. Cardiorespiratory and all-cause mortality after restrictions on sulphur content of fuel in Hong Kong: an intervention study. Lancet. 2002 Nov 23;360(9346):1646-52.
S13. Murray CJ, Nelson CR. State-space modeling of the relationship between air quality and mortality. J Air Waste Manag Assoc. 2000;50:1075-1080.
S14. Murray CJ, Lipfert FW. Inferring frail life expectancies in Chicago from daily fluctuations in elderly mortality. Inhal Toxicol. 2013;25:461-479.
S15. Knudsen SJ, Air pollution data for Toronto, Canada: A frail population state-space model, in New Trends in Statistical Modelling, Proc 16th Int Workshop on Statistical Modelling, Odense, Denmark, July 2-6, 2001. B Klein and L Korsholm, editors.
S16 Holford TR, Meza R, Warner KE, Meernik C, Jeon J, Moolgavkar SH, Levy DT. Tobacco control and the reduction in smoking-related premature deaths in the United States, 1964-2012. JAMA. 2014 Jan 8;311(2):164-71.
Additional Figures
Figure S1. Trends in county-level life expectancies are compared, showing similar dips ca. 1985-90 in New York, San Francisco, and New Orleans that may relate to the AIDS epidemic. This dip is not seen in Salt Lake City or Minneapolis, which may have had different lifestyles. This anomaly complicates cross-county comparisons and shows the influence of non-environmental factors. The range in life expectancies among these counties is 6-8 y.
Figure S2. The Harvard Six Cities Cohort Study began in the mid-1970s and has figured prominently in the air pollution – health effects literature. However, life expectancy calculations are limited to populations and cannot readily be computed for a mixed-age cohort; life expectancies for the six relevant counties are shown above. The relatively “clean” counties for Watertown and Topeka show steadily increasing life expectancies, which level off after about 1985 in the more polluted counties and in Portage, WI. The lower values in St. Louis may be due to its racial mix. All counties show steady increases in the 1970s, when the Clean Air Act took effect, regardless of their individual pollution levels. After about 1980, the relative rankings among cities vary. The range in life expectancies is 8-10 y.
.
Figure S3. There is a close correlation between mean life expectancies at birth and at age 65 (R=0.98, slope = 0.65 [sdev=0.03]). Thus, regression coefficients with respect to life expectancies at birth may be converted to those at age 65 by multiplying by 0.65. The intercept of this regression is 49; thus life expectancy at birth would be 49 when life expectancy at age 65 is 0. When life expectancy at birth is 100, life expectancy at age 65would be 32, so that the predicted mean attained age would be 98, which is reasonable agreement.
Figure S4 The standard deviation of life expectancies for the 3116 counties is a measure of their diversity. For life expectancy at birth, it reaches a minimum about 1983 and then increased back to its former level. The change in 99% confidence intervals is from about 10 –12 y. This suggests that counties tend become slightly less diverse and then become more diverse in later years because life expectancies improved greatly in some counties and leveled off in others. Initial indications are that this increase continues in subsequent years..
Figure S5. Life expectancies for males increase monotonically with decreased smoking but level off for females. The curves for males and females do not converge for either no-smokers or all-smokers, suggesting effects of other correlated variables. They imply an average loss of about 20 y per smoker, which is greater than has reported from direct studies by smoking habit. The difference between males and females is 5-6 y .
Figure S6. Relationships between life expectancy at age 65 and prevalence of smoking (shown as the percentage of non-smokers) are linear for males but tend to level off for females, similarly to Figure S5. Life expectancies for the two genders appear to converge at about 8 y for all-smokers and 19 y for no smokers. Thie regressions indicates a loss of about 9 y per smoker, even after age 65. However, these comparisons do not account for former smokers or the duration of smoking. It is also possible that other correlated behaviors contribute to the loss of life expectancy.
Figure S7. Life expectancies by state are strongly correlated with adult smoking prevalence in that state. Possible outliers are District of Columbia, which is 100% urban, and Utah is proscribed by religion as are other lifestyle choices. The correlations are 0.91 and 0.81 for birth and age 65 respectively. The slopes are –0.24 and –0.14 and are similar to the temporal slopes (-0.21,-0.09), which lends credence to both. However, these cross-state relationships are not matched by gender or time period.
Figure S8. Holford et al. (2014) show a linear decrease in smoking prevalence since 1960. Other data show peak smoking rates in the mid-1940s, perhaps because of military service and aggressive marketing. The mean rate of smoking has dropped by about 50% since the 1960s.